*Generate table of predicted versus actual attacks*
*Use "FPA Out Of Sample repdata.dta" for this figure*
*Incidents
logit      pirincidents_dum  gdp2006 sfi2006 ucdp2006 wto2006 polity2006 cinc2006 coast_ln pop2006  t1inc2006 t2inc2006 t3inc2006 if year>2006, cluster(ccode)
predict pinc
list ccode country year pirincidents_dum pinc if pinc>.5 &pinc!=. &year!=2006
list ccode country year pirincidents_dum pinc if pinc<.5 &pinc!=.&pirincidents_dum==1 &year!=2006

*Hijackings*
logit      hijack_dum gdp2006 sfi2006 ucdp2006 wto2006 polity2006 cinc2006 coast_ln pop2006  t1hij2006 t2hij2006 t3hij2006 if year>2006, cluster(ccode)
predict phij
list ccode country year hijack_dum phij if phij>.5 &phij!=. &year!=2006
list ccode country year hijack_dum phij if phij<.5 &phij!=.&hijack_dum==1 &year!=2006
